The broad objectives of this application are to determine the long-term course and outcome for schizophrenia and other severe and prolonged mental illness, and to assess the relative impact of individual premorbid and clinical factors, policy, and program variables on long-term course and outcome for persons with mental illness. The specific aims are: (1) to determine the impact of mental health service--system evolution in both Maine and Vermont, including policy, program, fiscal, and legal changes on long-term course and outcome for two matched cohorts (N=538); (2) to develop and construct empirically-derived patient typologies of the cohorts in both states using premorbid and symptom data, and to compare the longitudinal course trajectories and global outcomes for these typologies, and to compare the course and outcome of these empirically-derived typologies with other typologies from the literature on schizophrenia; and (3) to determine the longitudinal evolution of symptoms and symptom clusters for both cohorts through changes in the frequency of various symptoms over the lives of cohort members, and to study changes in the frequency of positive and negative symptoms across hospital admissions. Longitudinal analysis of symptom, policy, program, fiscal, and legal independent variables on various outcome domains in aim (1) will be accomplished through the use of multivariate linear regression techniques. Construction of typologies for aim (2) will involve the use of factor analytic and clustering methods. Aim (3) will involve the longitudinal frequency analysis of symptoms and a comparison of adjacent time periods.